The complexity of logic program updates

Research output: Chapter in Book / Conference PaperConference Paperpeer-review

1 Citation (Scopus)

Abstract

In the context of logic program updates, a knowledge base, which is presented as a logic program, can be updated in terms of another logic program, i.e. a set of update rules. In this paper, we investigate the complexity of logic program updates where conflict resolution on defeasible information is explicitly taken into account in an update. We show that in general the problem of model checking in logic program updates is co-NP-complete, and the corresponding inference problem is ΠP 2 -complete. We also characterize particular classes of update specifications where the inference problem has a lower computational complexity. These results confirm that logic program update, even if with the issue of conflict resolution on defeasible information to be presented, is not harder than the principal update tasks.

Original languageEnglish
Title of host publicationAI 2001
Subtitle of host publicationAdvances in Artificial Intelligence - 14th Australian Joint Conference on Artificial Intelligence, Proceedings
EditorsMarkus Stumptner, Dan Corbett, Mike Brooks, Dan Corbett
PublisherSpringer Verlag
Pages631-642
Number of pages12
ISBN (Print)9783540429609
DOIs
Publication statusPublished - 2001
Event14th Australian Joint Conference on Artificial Intelligence, AI 2001 - Adelaide, Australia
Duration: 10 Dec 200114 Dec 2001

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume2256
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference14th Australian Joint Conference on Artificial Intelligence, AI 2001
Country/TerritoryAustralia
CityAdelaide
Period10/12/0114/12/01

Bibliographical note

Publisher Copyright:
© Springer-Verlag Berlin Heidelberg 2001.

Fingerprint

Dive into the research topics of 'The complexity of logic program updates'. Together they form a unique fingerprint.

Cite this